OpenAdaptBook a call

For referral intake, claims, and the AI agents that do them

Do the task once. OpenAdapt does it every time.

A person, or a computer-use agent, does the task once in your EMR, claims system or portal. OpenAdapt turns that run into a program that makes no model calls. Before each save it confirms the right record, after it reads the record back, and when something changes it stops for a person instead of guessing.

Watch the demoBook a 30-minute call
Real run, 14 Oct 2026, 4x speed. Synthetic patient in a test copy of OpenEMR.

Watch it step by step →

Show it once. It runs as a program.

  1. 01

    Show the task

    A person does the task once while OpenAdapt records each click and field. A computer-use agent can show it instead.

    OpenEMR's Add Patient form with title, name and date of birth typed in by hand
    1. clickTitle
    2. typeMs.
    3. clickFirst name
    4. typeNadia
    5. clickLast name
    6. typeBrennan
  2. 02

    It becomes a program

    The recording compiles into fixed steps that run once per referral in your list. No AI model decides anything. A program has as many steps as the task; this OpenEMR referral has 41.

    This referral program · 41 steps

    next referrallist donechart notconfirmedLOOPFor each referralin the list01–09Name9 steps · 3 fields10–15Date of birth and sex6 steps · 2 fields16–29Address14 steps · 4 fields30–38Phone, email and country9 steps · 3 fields39–41Create the chart3 steps · checks the saved chartFinishedStop the runnext referrallist donechart notconfirmedLOOPFor each referralin the list01–09Name9 steps · 3 fields10–15Date of birth and sex6 steps · 2 fields16–29Address14 steps · 4 fields30–38Phone, email and country9 steps · 3 fields39–41Create the chart3 steps · checks the saved chartFinishedStop the run

    Compiled by OpenAdapt from one recorded referral on synthetic data, then set to run once per referral in a list.

  3. 03

    Every save is checked

    After saving, it reads the record back from the system, not the screen. By default a person approves each save.

    Chart read back · 12 of 12 match

    Title
    Ms.
    First name
    Nadia
    Last name
    Brennan
    Date of birth
    1971-04-19
    Sex
    Female
    Street
    27 WILLOW BEND ROAD
    City
    EXAMPLETON
    State
    MA

What OpenAdapt adds to your agent

Catches a bad saveA save that didn't land, caught.In a real quickstart run, the server rejected a note. The screen looked the same as a good save; the record check found no note, so it stopped.Asks a personApprove each save from your phone.One card per save, with a picture of exactly what will be saved. Swipe to save or stop, with 5 seconds to undo.No model on every entryThe same path every time. No screens sent, $0 an entry.
MeasureAgent aloneWith OpenAdapt
Model fees per referralabout $0.08$0
Correct saves62 of 6362 of 63
Median time36.5 s29.7 s
Timed referral runs in a test copy of OpenEMR, 14 to 15 Oct 2026, against Claude Sonnet 5.5 (claude-sonnet-5-5). Without page-structure identity the same task took 38.6 s, slower than the agent.

Who it's for

Healthcare and insurance operations

Referral intake, new-patient registration, claims intake and coverage look-ups in systems with no usable API. It runs on a computer you control, signed in to your system, and patient data stays there.

Healthcare →Insurance →Security →

Teams building AI agents

Your agent does the task once; OpenAdapt turns its run into a program and makes the entry from then on, with the right-record check, the read-back and a person's approval.

Describe the task, not a record. Don't enter patient data or personal information, such as names, dates of birth, or phone or record numbers. You get a page with a brief for your AI and the install steps. We keep the job text to see what people ask for.

For AI agent teams →Run it yourself →

Start with one workflow your team types every day.

  • Local Free, on your own computer
  • Team $499 a month
  • Enterprise Qualification sprints from $15k
Book a 30-minute callSee pricing →Sign in →